1927] Pilsbry-Bequaert, The Aquatic Mollusks of the Belgian Congo 559 
Mollusks are more frequently met with on the stretches with sandy 
or gravelly bottom just below or above the rocky barriers, especially in 
certain of the tributaries. Here one must look for the elegantly sculp- 
tured species of Melanoides, which live sometimes by the hundreds on the 
immersed edges of the sand spits. These little snails crawl over the soft, 
moist sand or more often are buried just beneath the semi-liquid surface 
of the bank. In a similar habitat are found a number of Unionide and 
Mutelide, the most striking being Mutela (Chelidonopsis) hirundo, 
appropriately named for the swallow-tail-shaped expansions of its 
posterior ends. It is still worthy of notice that in such areas certain 
sand-spits are abundantly provided with melanians and naiades, while 
many other, apparently similar banks may be examined in vain. There 
must be, of course, some ecological reason for these differences, but at 
present we have no satisfactory explanation to suggest. , 
An important feature of all these rivers are the seasonal variations 
In level to which they are subject. How far-reaching their effects are, 
even in the smaller tributaries, may be easily gathered from a comparison 
of the two photographs of Pl. LX VII, both representing the Garamba 
River in the country north of Faradje. That of Fig. 1 was taken in 
September, 1911, that is, toward the close of the rainy season, when the 
water reaches its highest level. Fig. 2 shows conditions as they were in 
February, 1913, about the middle of the dry season, when the water is 
quite low. The type specimens of the unionid Cexlatura mesafricana and 
of the mutelid Mutela garambx were found buried in the sand banks 
shown in Fig. 2. Similar conditions are illustrated for the Dungu River, 
at Faradje, in Pl. LXVIII, the picture representing conditions in the 
dry season (March, 1910) when Parreysia leopoldvillensis and Aspa- 
tharia sinuata were obtained in large numbers from the sand spits. 
The variations in level are more considerable in the main river 
Congo between Stanleyville and Leopoldville, where there are in addition 
two periods of flood every year. On the whole, the annual rise of the 
waters is very irregular, being sometimes more than one-third larger in 
succeeding years, as shown by the following figures': 
1These figures are based upon observations made during 1910, 1911, and 1912 and are taken from a 
short paper by H. Roussilhe, 1912. ‘Le régime des crues dans le réseau fluvial congolais.’ C. R. Ac. Sci. 
Paris, CLV, pp. 1141-1144. grr 
This author argues that the double annual flood of the Congo at Leopoldville is not due, as is gen- 
erally believed, to the combination of high water levels in the two hemispheres, resulting from the re- 
‘versed rainy and dry seasons north and south of the equator. He rather incriminates a double pluvio- 
metric annual cycle in the region of the sources of the southern headwaters where, he says, the 
‘tropical southern climate comprises two seasons of rains separated by a short dry season.’’ There ap- 
pears to be little evidence of this in the meteorological data published for the Katanga. 
